Machine learning in preclinical drug discovery.

Drug-discovery and drug-development endeavors are laborious, costly and time consuming. These programs can take upward of 12 years and cost US $2.5 billion, with a failure rate of more than 90%. Machine learning (ML) presents an opportunity to improve the drug-discovery process. Indeed, with the growing abundance of public and private large-scale biological and chemical datasets, ML techniques are becoming well positioned as useful tools that can augment the traditional drug-development process. In this Perspective, we discuss the integration of algorithmic methods throughout the preclinical phases of drug discovery. Specifically, we highlight an array of ML-based efforts, across diverse disease areas, to accelerate initial hit discovery, mechanism-of-action (MOA) elucidation and chemical property optimization. With advances in the application of ML across diverse therapeutic areas, we posit that fully ML-integrated drug-discovery pipelines will define the future of drug-development programs.

Discovery of antibiotics that selectively kill metabolically dormant bacteria.

There is a need to discover and develop non-toxic antibiotics that are effective against metabolically dormant bacteria, which underlie chronic infections and promote antibiotic resistance. Traditional antibiotic discovery has historically favored compounds effective against actively metabolizing cells, a property that is not predictive of efficacy in metabolically inactive contexts. Here, we combine a stationary-phase screening method with deep learning-powered virtual screens and toxicity filtering to discover compounds with lethality against metabolically dormant bacteria and favorable toxicity profiles. The most potent and structurally distinct compound without any obvious mechanistic liability was semapimod, an anti-inflammatory drug effective against stationary-phase E. coli and A. baumannii. Integrating microbiological assays, biochemical measurements, and single-cell microscopy, we show that semapimod selectively disrupts and permeabilizes the bacterial outer membrane by binding lipopolysaccharide. This work illustrates the value of harnessing non-traditional screening methods and deep learning models to identify non-toxic antibacterial compounds that are effective in infection-relevant contexts.

Discovery of a structural class of antibiotics with explainable deep learning.

The discovery of novel structural classes of antibiotics is urgently needed to address the ongoing antibiotic resistance crisis1-9. Deep learning approaches have aided in exploring chemical spaces1,10-15; these typically use black box models and do not provide chemical insights. Here we reasoned that the chemical substructures associated with antibiotic activity learned by neural network models can be identified and used to predict structural classes of antibiotics. We tested this hypothesis by developing an explainable, substructure-based approach for the efficient, deep learning-guided exploration of chemical spaces. We determined the antibiotic activities and human cell cytotoxicity profiles of 39,312 compounds and applied ensembles of graph neural networks to predict antibiotic activity and cytotoxicity for 12,076,365 compounds. Using explainable graph algorithms, we identified substructure-based rationales for compounds with high predicted antibiotic activity and low predicted cytotoxicity. We empirically tested 283 compounds and found that compounds exhibiting antibiotic activity against Staphylococcus aureus were enriched in putative structural classes arising from rationales. Of these structural classes of compounds, one is selective against methicillin-resistant S. aureus (MRSA) and vancomycin-resistant enterococci, evades substantial resistance, and reduces bacterial titres in mouse models of MRSA skin and systemic thigh infection. Our approach enables the deep learning-guided discovery of structural classes of antibiotics and demonstrates that machine learning models in drug discovery can be explainable, providing insights into the chemical substructures that underlie selective antibiotic activity.

Applications of machine learning in microbial natural product drug discovery.

INTRODUCTION: Natural products (NPs) are a desirable source of new therapeutics due to their structural diversity and evolutionarily optimized bioactivities. NPs and their derivatives account for roughly 70% of approved pharmaceuticals. However, the rate at which novel NPs are discovered has decreased. To accelerate the microbial NP discovery process, machine learning (ML) is being applied to numerous areas of NP discovery and development. AREAS COVERED: This review explores the utility of ML at various phases of the microbial NP drug discovery pipeline, discussing concrete examples throughout each major phase: genome mining, dereplication, and biological target prediction. Moreover, the authors discuss how ML approaches can be applied to semi-synthetic approaches to drug discovery. EXPERT OPINION: Despite the important role that microbial NPs play in the development of novel drugs, their discovery has declined due to challenges associated with the conventional discovery process. ML is positioned to overcome these limitations given its ability to model complex datasets and generalize to novel chemical and sequence space. Unsurprisingly, ML comes with its own limitations that must be considered for its successful implementation. The authors stress the importance of continuing to build high quality and open access NP datasets to further increase the utility of ML in NP discovery.

Deep learning-guided discovery of an antibiotic targeting Acinetobacter baumannii.

Acinetobacter baumannii is a nosocomial Gram-negative pathogen that often displays multidrug resistance. Discovering new antibiotics against A. baumannii has proven challenging through conventional screening approaches. Fortunately, machine learning methods allow for the rapid exploration of chemical space, increasing the probability of discovering new antibacterial molecules. Here we screened ~7,500 molecules for those that inhibited the growth of A. baumannii in vitro. We trained a neural network with this growth inhibition dataset and performed in silico predictions for structurally new molecules with activity against A. baumannii. Through this approach, we discovered abaucin, an antibacterial compound with narrow-spectrum activity against A. baumannii. Further investigations revealed that abaucin perturbs lipoprotein trafficking through a mechanism involving LolE. Moreover, abaucin could control an A. baumannii infection in a mouse wound model. This work highlights the utility of machine learning in antibiotic discovery and describes a promising lead with targeted activity against a challenging Gram-negative pathogen.

Antibiotic discovery in the artificial intelligence era.

As the global burden of antibiotic resistance continues to grow, creative approaches to antibiotic discovery are needed to accelerate the development of novel medicines. A rapidly progressing computational revolution-artificial intelligence-offers an optimistic path forward due to its ability to alleviate bottlenecks in the antibiotic discovery pipeline. In this review, we discuss how advancements in artificial intelligence are reinvigorating the adoption of past antibiotic discovery models-namely natural product exploration and small molecule screening. We then explore the application of contemporary machine learning approaches to emerging areas of antibiotic discovery, including antibacterial systems biology, drug combination development, antimicrobial peptide discovery, and mechanism of action prediction. Lastly, we propose a call to action for open access of high-quality screening datasets and interdisciplinary collaboration to accelerate the rate at which machine learning models can be trained and new antibiotic drugs can be developed.

Reactive metabolic byproducts contribute to antibiotic lethality under anaerobic conditions.

Understanding how bactericidal antibiotics kill bacteria remains an open question. Previous work has proposed that primary drug-target corruption leads to increased energetic demands, resulting in the generation of reactive metabolic byproducts (RMBs), particularly reactive oxygen species, that contribute to antibiotic-induced cell death. Studies have challenged this hypothesis by pointing to antibiotic lethality under anaerobic conditions. Here, we show that treatment of Escherichia coli with bactericidal antibiotics under anaerobic conditions leads to changes in the intracellular concentrations of central carbon metabolites, as well as the production of RMBs, particularly reactive electrophilic species (RES). We show that antibiotic treatment results in DNA double-strand breaks and membrane damage and demonstrate that antibiotic lethality under anaerobic conditions can be decreased by RMB scavengers, which reduce RES accumulation and mitigate associated macromolecular damage. This work indicates that RMBs, generated in response to antibiotic-induced energetic demands, contribute in part to antibiotic lethality under anaerobic conditions.

A brief guide to machine learning for antibiotic discovery.

Rising antibiotic resistance and an alarmingly lean antibiotic pipeline require the adoption of novel approaches to rapidly discover new structural and functional classes of antibiotics. Excitingly, algorithmic approaches to antibiotic discovery are sufficiently advanced to meaningfully influence the antibiotic discovery process. Indeed, once trained on high-quality datasets, contemporary machine-learning and deep-learning models can be used to perform predictions for new antibiotics across vast chemical spaces, orders of magnitude more rapidly than compounds can be screened in the laboratory. This increases the probability of discovering new antibiotics with desirable properties. In this short review, we briefly describe the utility of contemporary machine-learning and deep-learning approaches to guide the discovery of new small-molecule antibiotics and unidentified natural products. We then propose a call to action for more open sharing of high-quality screening datasets to accelerate the rate at which forthcoming antibiotic-prediction models can be trained. Together, we aim to introduce antibiotic discoverers to a sample of recent applications of contemporary algorithmic methods to facilitate the wider adoption of these powerful computational approaches.

Deep learning identifies synergistic drug combinations for treating COVID-19.

Effective treatments for COVID-19 are urgently needed. However, discovering single-agent therapies with activity against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been challenging. Combination therapies play an important role in antiviral therapies, due to their improved efficacy and reduced toxicity. Recent approaches have applied deep learning to identify synergistic drug combinations for diseases with vast preexisting datasets, but these are not applicable to new diseases with limited combination data, such as COVID-19. Given that drug synergy often occurs through inhibition of discrete biological targets, here we propose a neural network architecture that jointly learns drug-target interaction and drug-drug synergy. The model consists of two parts: a drug-target interaction module and a target-disease association module. This design enables the model to utilize drug-target interaction data and single-agent antiviral activity data, in addition to available drug-drug combination datasets, which may be small in nature. By incorporating additional biological information, our model performs significantly better in synergy prediction accuracy than previous methods with limited drug combination training data. We empirically validated our model predictions and discovered two drug combinations, remdesivir and reserpine as well as remdesivir and IQ-1S, which display strong antiviral SARS-CoV-2 synergy in vitro. Our approach, which was applied here to address the urgent threat of COVID-19, can be readily extended to other diseases for which a dearth of chemical-chemical combination data exists.

Cytoplasmic condensation induced by membrane damage is associated with antibiotic lethality.

Bactericidal antibiotics kill bacteria by perturbing various cellular targets and processes. Disruption of the primary antibiotic-binding partner induces a cascade of molecular events, leading to overproduction of reactive metabolic by-products. It remains unclear, however, how these molecular events contribute to bacterial cell death. Here, we take a single-cell physical biology approach to probe antibiotic function. We show that aminoglycosides and fluoroquinolones induce cytoplasmic condensation through membrane damage and subsequent outflow of cytoplasmic contents as part of their lethality. A quantitative model of membrane damage and cytoplasmic leakage indicates that a small number of nanometer-scale membrane defects in a single bacterium can give rise to the cellular-scale phenotype of cytoplasmic condensation. Furthermore, cytoplasmic condensation is associated with the accumulation of reactive metabolic by-products and lipid peroxidation, and pretreatment of cells with the antioxidant glutathione attenuates cytoplasmic condensation and cell death. Our work expands our understanding of the downstream molecular events that are associated with antibiotic lethality, revealing cytoplasmic condensation as a phenotypic feature of antibiotic-induced bacterial cell death.

Clinically relevant mutations in core metabolic genes confer antibiotic resistance.

Although metabolism plays an active role in antibiotic lethality, antibiotic resistance is generally associated with drug target modification, enzymatic inactivation, and/or transport rather than metabolic processes. Evolution experiments of Escherichia coli rely on growth-dependent selection, which may provide a limited view of the antibiotic resistance landscape. We sequenced and analyzed E. coli adapted to representative antibiotics at increasingly heightened metabolic states. This revealed various underappreciated noncanonical genes, such as those related to central carbon and energy metabolism, which are implicated in antibiotic resistance. These metabolic alterations lead to lower basal respiration, which prevents antibiotic-mediated induction of tricarboxylic acid cycle activity, thus avoiding metabolic toxicity and minimizing drug lethality. Several of the identified metabolism-specific mutations are overrepresented in the genomes of >3500 clinical E. coli pathogens, indicating clinical relevance.

Eradicating Bacterial Persisters with Combinations of Strongly and Weakly Metabolism-Dependent Antibiotics.

The vast majority of bactericidal antibiotics display poor efficacy against bacterial persisters, cells that are in a metabolically repressed state. Molecules that retain their bactericidal functions against such bacteria often display toxicity to human cells, which limits treatment options for infections caused by persisters. Here, we leverage insight into metabolism-dependent bactericidal antibiotic efficacy to design antibiotic combinations that sterilize both metabolically active and persister cells, while minimizing the antibiotic concentrations required. These rationally designed antibiotic combinations have the potential to improve treatments for chronic and recurrent infections.

A Deep Learning Approach to Antibiotic Discovery.

Due to the rapid emergence of antibiotic-resistant bacteria, there is a growing need to discover new antibiotics. To address this challenge, we trained a deep neural network capable of predicting molecules with antibacterial activity. We performed predictions on multiple chemical libraries and discovered a molecule from the Drug Repurposing Hub-halicin-that is structurally divergent from conventional antibiotics and displays bactericidal activity against a wide phylogenetic spectrum of pathogens including Mycobacterium tuberculosis and carbapenem-resistant Enterobacteriaceae. Halicin also effectively treated Clostridioides difficile and pan-resistant Acinetobacter baumannii infections in murine models. Additionally, from a discrete set of 23 empirically tested predictions from >107 million molecules curated from the ZINC15 database, our model identified eight antibacterial compounds that are structurally distant from known antibiotics. This work highlights the utility of deep learning approaches to expand our antibiotic arsenal through the discovery of structurally distinct antibacterial molecules.

Bacterial metabolic state more accurately predicts antibiotic lethality than growth rate.

Growth rate and metabolic state of bacteria have been separately shown to affect antibiotic efficacy1-3. However, the two are interrelated as bacterial growth inherently imposes a metabolic burden4; thus, determining individual contributions from each is challenging5,6. Indeed, faster growth is often correlated with increased antibiotic efficacy7,8; however, the concurrent role of metabolism in that relationship has not been well characterized. As a result, a clear understanding of the interdependence between growth and metabolism, and their implications for antibiotic efficacy, are lacking9. Here, we measured growth and metabolism in parallel across a broad range of coupled and uncoupled conditions to determine their relative contribution to antibiotic lethality. We show that when growth and metabolism are uncoupled, antibiotic lethality uniformly depends on the bacterial metabolic state at the time of treatment, rather than growth rate. We further reveal a critical metabolic threshold below which antibiotic lethality is negligible. These findings were general for a wide range of conditions, including nine representative bactericidal drugs and a diverse range of Gram-positive and Gram-negative species (Escherichia coli, Acinetobacter baumannii and Staphylococcus aureus). This study provides a cohesive metabolic-dependent basis for antibiotic-mediated cell death, with implications for current treatment strategies and future drug development.

Bacterial Metabolism and Antibiotic Efficacy.

Antibiotics target energy-consuming processes. As such, perturbations to bacterial metabolic homeostasis are significant consequences of treatment. Here, we describe three postulates that collectively define antibiotic efficacy in the context of bacterial metabolism: (1) antibiotics alter the metabolic state of bacteria, which contributes to the resulting death or stasis; (2) the metabolic state of bacteria influences their susceptibility to antibiotics; and (3) antibiotic efficacy can be enhanced by altering the metabolic state of bacteria. Altogether, we aim to emphasize the close relationship between bacterial metabolism and antibiotic efficacy as well as propose areas of exploration to develop novel antibiotics that optimally exploit bacterial metabolic networks.

A multiplexable assay for screening antibiotic lethality against drug-tolerant bacteria.

Antibiotic screens typically rely on growth inhibition to characterize compound bioactivity-an approach that cannot be used to assess the bactericidal activity of antibiotics against bacteria in drug-tolerant states. To address this limitation, we developed a multiplexed assay that uses metabolism-sensitive staining to report on the killing of antibiotic-tolerant bacteria. This method can be used with diverse bacterial species and applied to genome-scale investigations to identify therapeutic targets against tolerant pathogens.

Our Evolving Understanding of the Mechanism of Quinolones.

The maintenance of DNA supercoiling is essential for the proper regulation of a plethora of biological processes. As a consequence of this mode of regulation, ahead of the replication fork, DNA replication machinery is prone to introducing supercoiled regions into the DNA double helix. Resolution of DNA supercoiling is essential to maintain DNA replication rates that are amenable to life. This resolution is handled by evolutionarily conserved enzymes known as topoisomerases. The activity of topoisomerases is essential, and therefore constitutes a prime candidate for targeting by antibiotics. In this review, we present hallmark investigations describing the mode of action of quinolones, one of the antibacterial classes targeting the function of topoisomerases in bacteria. By chronologically analyzing data gathered on the mode of action of this imperative antibiotic class, we highlight the necessity to look beyond primary drug-target interactions towards thoroughly understanding the mechanism of quinolones at the level of the cell.

Overcoming mcr-1 mediated colistin resistance with colistin in combination with other antibiotics.

Plasmid-borne colistin resistance mediated by mcr-1 may contribute to the dissemination of pan-resistant Gram-negative bacteria. Here, we show that mcr-1 confers resistance to colistin-induced lysis and bacterial cell death, but provides minimal protection from the ability of colistin to disrupt the Gram-negative outer membrane. Indeed, for colistin-resistant strains of Enterobacteriaceae expressing plasmid-borne mcr-1, clinically relevant concentrations of colistin potentiate the action of antibiotics that, by themselves, are not active against Gram-negative bacteria. The result is that several antibiotics, in combination with colistin, display growth-inhibition at levels below their corresponding clinical breakpoints. Furthermore, colistin and clarithromycin combination therapy displays efficacy against mcr-1-positive Klebsiella pneumoniae in murine thigh and bacteremia infection models at clinically relevant doses. Altogether, these data suggest that the use of colistin in combination with antibiotics that are typically active against Gram-positive bacteria poses a viable therapeutic alternative for highly drug-resistant Gram-negative pathogens expressing mcr-1.

Bicarbonate Alters Bacterial Susceptibility to Antibiotics by Targeting the Proton Motive Force.

The antibacterial properties of sodium bicarbonate have been known for years, yet the molecular understanding of its mechanism of action is still lacking. Utilizing chemical-chemical combinations, we first explored the effect of bicarbonate on the activity of conventional antibiotics to infer on the mechanism. Remarkably, the activity of 8 classes of antibiotics differed in the presence of this ubiquitous buffer. These interactions and a study of mechanism of action revealed that, at physiological concentrations, bicarbonate is a selective dissipater of the pH gradient of the proton motive force across the cytoplasmic membrane of both Gram-negative and Gram-positive bacteria. Further, while components that make up innate immunity have been extensively studied, a link to bicarbonate, the dominant buffer in the extracellular fluid, has never been made. Here, we also explored the effects of bicarbonate on components of innate immunity. Although the immune response and the buffering system have distinct functions in the body, we posit there is interplay between these, as the antimicrobial properties of several components of innate immunity were enhanced by a physiological concentration of bicarbonate. Our findings implicate bicarbonate as an overlooked potentiator of host immunity in the defense against pathogens. Overall, the unique mechanism of action of bicarbonate has far-reaching and predictable effects on the activity of innate immune components and antibiotics. We conclude that bicarbonate has remarkable power as an antibiotic adjuvant and suggest that there is great potential to exploit this activity in the discovery and development of new antibacterial drugs by leveraging testing paradigms that better reflect the physiological concentration of bicarbonate.